Construction and Comparison of Sound Quality Index for the Vehicle HVAC System Using Regression Model and Neural Network Model

The reduction of the vehicle interior noise has been the main interest of noise and vibration harshness (NVH) engineers. The driver's perception on the vehicle noise is affected largely by psychoacoustic characteristic of the noise as well as the SPL. In particular, the heating, ventilation and air conditioning (HVAC) system sound among the vehicle interior noise has been reflected sensitively in psychoacoustics view point. Even though the HVAC noise is not louder than overall noise level, it clearly affects subjective perception to drivers in the way of making to be nervous or annoyed. Therefore, these days a vehicle engineer takes aim at developing sound quality as well as reduction of noise. In this paper, we acquired noises in the HVAC from many vehicles. Through the objective and subjective sound quality (SQ) evaluation with acquiring noises recorded by the vehicle HVAC system, the simple and multiple regression models were obtained for the subjective evaluation 'Pleasant' using the semantic differential method (SDM). The regression procedure also allows you to produce diagnostic statistics to evaluate the regression estimates including appropriation and accuracy. Furthermore, the neural network (NN) model were obtained using three inputs(loudness, sharpness and roughness) of the SQ metrics and one output(subjective 'Pleasant'). Because human's perception is very complex and hard to estimate their pattern, we used NN model. The estimated models were compared with correlations between output indexes of SQ and hearing test results for verification data 'Pleasant'. As a result of application of the SQ indexes, the NN model was shown with the largest correlation of SQ indexes and we found possibilities to predict the SQ metrics.